AI in Finance Summit New York schedule

•    Protecting Client Data in the age of AI
•    A safe AI execution loop
•    Metadata Driven practical solution

 

 

•    Enhance risk detection with LLMs that learn from historical investigations and analyst notes.
•    Empower investigation teams with AI copilots for instant insights, narratives, and recommended actions.
•    Automate AML/KYC processes using explainable, auditable AI for transparent compliance.
•    Strengthen defenses by leveraging AI to proactively identify emerging threats and reduce noise

 

 

Can we open the black box of large language models and make AI in finance truly transparent? As financial institutions adopt LLMs for tasks like trading, compliance, and advisory, the need for transparency is more critical than ever. Understanding how these models internally reason about financial topics is key to ensuring trust and regulatory alignment. This talk covers how mechanistic interpretability can reveal internal patterns in LLMs related and use it for applications such as trading, sentiment analysis and hallucination reduction.

As a business leader and technology provider Purnima Padmanabhan - General Manager of Broadcom’s, Tanzu Division - will share how the Tanzu engineering team’s experience of adopting AI coding assistants has informed the way we build software products and optimize Tanzu for agentic use cases that meet enterprise-scale requirements and security standards.  
In her talk Purnima will outline the success factors necessary to get beyond aspirational ambitions to truly transformative and practical use cases that will make a difference to your business. She’ll outline a proven approach to take you from experimentation and guided usage through to establishing an iterative and productive operating model, to delivering features and apps at scale, safely.  
 
Topics include: 
•    Agentic Runtime Considerations: Establishing guardrails, observability and standardization for autonomous AI agents operating in secure, regulated environments. 
•    Data & AI Sovereignty: Balancing regulatory compliance and local control with the flexibility of an open ecosystem of models, frameworks and data sources. 
•    Robust, Centralized AI Governance: Defining governance for agents such as coding assistants, and enforcing controls on agent access to sensitive enterprise tools and data 

•    What are the most effective strategies to scale AI initiatives while keeping costs under control?
•    How can organizations balance the transformative potential of generative AI with the risks it introduces?
•    In what ways AI Agents are reshaping business operations, and what challenges come with their adoption?
•    How do you measure and ensure tangible business value from AI investments across the enterprise?

Panelists:
Sreekar Bhaviripudi, Head of Machine Learning, MORGAN STANLEY
Micha Kiener, CTO, FLOWABLE
Raj Gunukula, Group Technical Program Manager, COINBASE
Alp Basol, Head of Artificial Intelligence, COBANK
Sai Zeng, Head of AI, Executive Director, Investment Banking & Global Capital Markets Technology, MORGAN STANLEY

Moderator: Judson Beaver, Client Solutions Architect, CORETEK

Financial institutions are entering a new stage of AI adoption in which model quality and data hygiene alone are insufficient. The main challenge is developing an architecture that can continuously provide data, context, and computation to production AI systems without compromising performance, control, or resilience. Most importantly, it must also meet regulatory requirements. 
This session explores how companies are creating AI environments that combine accelerated computing with high-throughput data platforms to support demanding financial workloads and efficiently coordinate data and GPU resources across teams and use cases. Participants will learn practical ways modern AI infrastructure can help financial firms reduce deployment times, improve resource utilization, and lay the foundation for secure, high-value monetization. 

 

•    We have spent years perfecting traditional Machine Learning models, that are world class at predicting risk or generating Alpha. From a business perspective, what is the tangible benefit of moving to an Agentic Workflow?
•    Finance is built on deterministic rules and traditional Machine Learning models “if X happens, Y must follow”. Yet, multi-agent systems are inherently probabilistic. How do you reconcile a probabilistic brain with a deterministic ledger when managing someone’s life savings? Are we just building more sophisticated black boxes that we cant truly audit until after they fail?
•    When agent move at millisecond speeds, the “human in the loop” can feel like an illusion.  How do we hard code the human as a supervisor role, so we dont lose the oversight we had with traditional ML?
•    In 2010, we had a Flash Crash caused by simple algorithms. Today, we face an ecosystem where millions of autonomous agents may interact in untested ways. Is our rush toward autonomous agents creating a new digital systemic risk where agents might spiral into market wide liquidation?
•    By the 2027 AI in Finance Summit in New York, what is one financial task currently performed by a human that will be 100% handled by an autonomous multi-agent system?

Panelists:
Michael Mocanu, Sr. Director, Data Science & AI, LIBERTY MUTUAL INSURANCE
Bijit Ghosh, Managing Director - AI/ML/Data, WELLS FARGO
Brij Mohan, Vice President-Software Development, LPL FINANCIAL
Jake Katz, Head of Analytics Research, LONDON STOCK EXCHANGE GROUP
Lily Li, Head of AI Adoption and Solutions, FRANKLIN TEMPLETON

Moderator: Dave Coluccio, Head of Buy Side Segment and Data Feed Strategy
S&P GLOBAL MARKET INTELLIGENCE

Most AI deployments in financial services never make it past the pilot phase. Dimitri shares six lessons from deploying AI agents that automate the full lifecycle of customer operations at production scale across leading banks and fintechs, covering:

• Why cost ROI on AI agents is more nuanced than it seems

• What it actually takes to go from 50% to 80%+ automation rate

• The biggest upside of AI agents in finance right now (it's not cost savings)

• How to get to production faster

• Why testing is the biggest blind spot in AI agent deployment

• The pitfalls of both building and buying, and why neither framing is quite right

• Evolution of quant analysis on Wall Street — a real-world scenario
• Balancing innovation, speed and risk
• Delivering business value while avoiding horror stories
• Where people, methods and technology meet

Why agentic AI gets stuck before production — and what's really slowing financial teams downHow to compare experiments and build a compliance-ready evidence trail before you shipA faster, cleaner path to production that brings engineering, risk, and compliance teams to the same page

•    Why cloud-only AI strategies in financial services are driving spiralling, unpredictable inference costs and increasing data privacy and sovereignty risk
•    How to use a hybrid AI approach spanning local, cloud, and on-premises, to run workloads based on cost, sensitivity, and latency, including local-first experimentation on workstation class hardware.
•    What architectural patterns leading institutions are adopting (hybrid AI data fabric and governed data platforms) to build AI systems that are economically sustainable, compliant, and resilient.
As financial institutions race to deploy AI across fraud, risk, trading, and customer experience, many are discovering that cloud-only AI strategies are driving spiraling, and often unpredictable, inference costs and creating new data privacy and sovereignty risks. This keynote reframes the conversation from “cloud vs data center” to ‘cloud and data center” – a modern hybrid AI approach spanning local, cloud, and on premises—and makes the business case for why on-prem must be a critical consideration for high-volume, sensitive, and latency critical workloads. Attendees will learn how to evaluate deployment choices when running AI workloads, understand the true economics of inference at scale, and see emerging architectural patterns from leading financial institutions, including local-first experimentation with desktop workstations (such as Dell Pro Precision plus NVIDIA DGX Spark with GB10 Grace Blackwell Superchip), a hybrid AI data fabric, and governed data platforms. The session closes with practical guidance on how to rebalance between cloud and on-prem to build AI systems that are economically sustainable, compliant, and resilient.

•    Agentic AI driving smarter trading decisions.
•    Multi-agent systems for better forecasting and risk.
•    Quantum computing reshaping compliance.
•    Strategic impact for financial institutions.

•    The Infrastructure Gap AI Finance Can't Ignore
    Financial institutions are already onchain: JPMorgan tokenizing treasuries, Stripe with stablecoin payments, Robinhood moving billions
    AI agents need to interact with these systems reliably, at scale, cost-effectively
    Just like you couldn't scrape websites fast enough to build Netflix, you can't hit public RPC nodes and build production financial AI
•    Why AWS Emerged (And Why Cortex Exists)
    HTTP existed, but building scalable applications was impossible without infrastructure abstraction
    Blockchain protocols exist, but building AI agents that interact with onchain finance is operationally infeasible without the right data layer
    Cortex solved for financial AI what AWS solved for cloud apps: reliability (99.9% uptime vs. node failures), speed (80% faster than competitors), cost (90% reduction), and abstractions AI agents actually need
•    What You Can Build Today (Live Demo)
    AI agents processing stablecoin payments autonomously
    Treasury management with tokenized assets
    Real-time compliance monitoring across onchain activity

•    How can we use real-time data streams to make better LLMs? (ex: fine tuning)
•    How can we use real-time data for better AI inferencing? (ex: observability & evaluation)
•    How can we use real-time data for agentic systems (RAG and MCP architectures)?
Documented here

•    Failures First – What early missteps revealed about AI’s real limits in finance.
•    Scaling Wins – How pioneers turned pilots into enterprise-wide impact.
•    Trust Factor – Building governance and transparency from the start.
•    Next Moves – What early adopters see as the boldest opportunities ahead

•    How does poor data quality directly undermine AI outcomes in finance?
•    What governance practices truly build trust in AI-driven decisions?
•    How can institutions balance innovation with strict data controls?
•    What lessons from early adopters show the ROI of strong data governance?

Panelists:
Tyler Frieling, Director, Emerging Technologies, BLACKROCK
Anupama Garani, AI & Machine Learning, PIMCO
Schitiz Saxena, Former Director - Chief Data Office, TD
Julia Cherashore, Senior Fellow, DATA FOUNDATION
Nishit Dhilen Mehta, Vice President, Data Analytics, JPMORGAN CHASE

Moderator: Michael Moore, Senior Director, Strategy and Innovation, Neo4j

Live demonstrations to explore how AI agents can be embedded into document/spreadsheet workflows in finance where intelligence becomes part of daily work (e.g., FP&A, reporting, modelling). Experience real-time interactions with AI agents across documents, data, and presentations, showing how a single agentic approach can be applied consistently across recurring, common themes
Build and customize simple AI agents during the session, giving participants first-hand experience of rapid agent development with practical guardrails, triggers, and approvals - without conventional engineering efforts.

•    Thriving in uncertainty, AI built to perform under market volatility and rapid change.
•    Trust by design, resilience rooted in transparency, governance, and ethical AI.
•    Future-ready finance, adaptable AI that meets new risks, regulations, and customer demands

•    By 2030, what will distinguish financial institutions that successfully harness AI from those that fall behind?
•    Which areas of the financial sector are most ripe for AI-driven transformation in the next five years?
•    How can we balance automation with transparency, fairness, and human oversight as AI becomes central to financial decision-making?
•    How will AI reshape the roles, skills, and culture of professionals in the finance industry by 2030?

Panelists:
Arjun Wadwalkar, Senior Manager of Product Management, GLOBAL PAYMENTS
Jake Katz, Head of Analytics Research, LONDON STOCK EXCHANGE GROUP
Matt Goldwasser, Head of AI Data Science, T. ROWE PRICE 
Aakanksha Jadhav, Director Product Development, MASTERCARD

Moderator: Impetus

 
 

As enterprises race toward widespread adoption of Generative AI and Agentic AI, many struggle to keep pace with the speed of innovation while maintaining governance, security, and operational rigor. This session explores how to design and operationalize an Enterprise AI Hub that enables rapid experimentation without sacrificing production readiness.
Drawing from the CAI AI Hub journey, this talk outlines a practical, real world approach to transforming AI delivery through self service engineering experiences, standardized blueprints, and a modular, open reference architecture. Attendees will learn how shifting from pipeline first development to a Run → Experiment → Preview → Promote model dramatically reduces friction for AI engineers and data scientists.
The session will cover:
•    How to abstract infrastructure complexity while preserving enterprise-grade controls
•    Designing self service AI platforms with reusable blueprints for RAG, agents, and multi agent orchestration
•    Enabling fast experimentation via playground and builder environments while supporting governed promotion to production
•    Embedding observability, cost controls, and Responsible AI by default
•    Lessons learned from rolling out GenAI and Agentic capabilities at scale across multiple teams and business domains
This talk is ideal for technology leaders, platform architects, and AI practitioners looking to move beyond isolated pilots and build AI platforms that scale innovation, accelerate time to value, and prepare organizations for the next wave of intelligent system

 

 

Challenge: As Generative and Agentic AI systems are increasingly used in high impact environments, validating their behaviour in a scalable and trustworthy way remains a major challenge.
Approach: We present an adaptive, panel based validation framework that dynamically selects and combines multiple evaluators based on context and reliability.
Outcome: This approach enables more robust, scalable, and context aware validation, improving confidence in AI systems deployed in critical settings.

 

 

•    How can financial institutions strike the right balance between rapid AI innovation and strict compliance requirements?
•    What are the most common security risks when deploying AI in finance, and how can organizations mitigate them?
•    How can firms ensure that AI tools are not only compliant at launch but remain compliant as regulations evolve?
•    What does it take to scale AI solutions across an enterprise without compromising governance or customer trust?
•    Looking ahead, what capabilities will define the next generation of AI tools that are truly enterprise-ready for finance

Dhivya Nagasubramanian, Vice President, AI Transformation & Innovation, U.S. BANK
Leah Price, General Manager, Tinman AI Platform, BETTER
Ellis Wong, Chief Information Security Officer, JST CAPITAL
Izge Cengiz Ercan, Director of Strategic Innovation, VALLEY BANK


Moderator: Nick Goble, Director of Solution Architecture, Financial Services and Insurance, DOMINO DATA LAB

Assess the shift from discrete AI agents to embedded, ambient intelligence within core finance and customer-facing systems, and present implications for scale, governance, and user adoption
Understand why human-in-the-loop design and event-driven triggers are critical for actionable, auditable, and trusted AI in regulated environments.
Highlight how new development paradigms enable business teams to design and deploy internal AI agents faster, unlocking speedy domain-led innovation while staying aligned with risk and compliance requirements.

Financial institutions are rapidly integrating AI into core business functions – from underwriting and customer engagement to compliance and internal decision-making. But the experimentation phase is over. As AI becomes embedded in daily operations, regulators and litigants are evaluating its use through existing disclosure, supervisory, and enforcement frameworks. The question is no longer whether institutions are using AI – it is whether they can demonstrate governance, defensibility, and compliance when challenged.

In this session, Kate Reilly and Aaron Miner, securities partners at international AmLaw 100 firm Arnold & Porter will offer a legal and enforcement-focused perspective on AI risk management in the financial sector. They will address preservation and discoverability of AI prompts and outputs, privilege risks highlighted in emerging decisions (including recent scrutiny of whether AI prompts may be protected), accountability for AI-generated errors, and exposure created by employee use of external or personal AI tools, and third party/vendor AI risk and data privacy or training data governance concerns.

The discussion will also cover disclosure obligations, supervisory expectations, and “Training 2.0” – practical governance measures institutions should implement now to reduce litigation and enforcement risk as AI adoption accelerates.

•    How can financial institutions ensure the quality, integrity, and governance of customer data to drive reliable AI outcomes?
•    In what ways does trusted customer data accelerate the development of innovative, personalized financial products and services?
•    What are the biggest challenges in balancing customer privacy with the need for data to fuel AI innovation?
•    How can organizations build and maintain customer trust while scaling AI solutions that rely heavily on sensitive data?

Panelists: 
Reema Gill, Data/AI Governance Specialist, WEALTHSIMPLE 
Shaurya Tripathi, Senior Data Scientist, MUNICH RE 
Prashant Reddy, Co-Founder & CEO, ARTIAN AI 
Santosh Gaikwad, Executive Director - Artificial Intelligence Platform Engineering Lead, JPMORGAN CHASE 

Moderator: Yatharth Sejpal, CEO, KNOWIDEA Technologies Ltd 
 

•    Beyond automation: how AI agents move from task execution to strategic decision-making
•    Efficiency unlocked: streamlining operations and reducing costs across the financial ecosystem
•    Revenue engine: transforming AI agents into drivers of growth, new products, and customer value
•    The future of finance: AI agents reshaping roles, talent, and business models

•    How attackers gain access with cutting edge attacks like simply asking for money to be transferred 
•    How cloud and AI fits into attack and defence 
•    How you can better defend you and your company from these attacks

•    From Pilots to Scalable Impact: How financial institutions can evolve AI maturity to unlock measurable business outcomes.
•    The Growth Catalyst: Why AI maturity is the key to driving efficiency, innovation, and revenue generation in financial services.
•    Building Trust at Scale: How to align governance, compliance, and customer trust as AI capabilities mature.

•    How to overcome the challenges of aligning business and engineering on the AI journey
•    Unlocking the power of process optimization to drive sustainable profitability
•    Strategies for building a bridge between process and delivery in ai finance
•    Best practices for implementing ai solutions that deliver real-world results 

Trust, transparency, and explainability as the foundation for AI adoption in financial services.
How evolving regulations are shaping responsible AI design and governance.
Turning ethical AI principles into practical, scalable implementation.
Balancing innovation with risk management and long-term compliance.

 

This session puts people at the center of AI explainability. Using real banking examples, it offers a practical framework to identify high-value AI use cases and design explanations that are clear, responsible, and trusted across the organization.

•    Objective of the use of AI - Productivity or doing things differently with AI/ How is GenAI different compared to traditional AI/ML and how can it be used in the business context? 
•    What would you recommend companies do to build AI successfully from the ground up? Is data still as critical in the new context of GenAI? 
•    Resistance to change, & elements to a successful AI Adoption? Harsh Singh, Manager, Data 
•    How are you using AI in your personal life?

•    From Reactive to Proactive: AI-powered analytics to predict and prevent fraud before it happens.
•    Real-Time Defense: Stopping fraudulent transactions instantly while protecting customer experience.
•    Agile & Adaptive Models: Continuously evolving with new fraud schemes and financial crime tactics.
•    Trust as Currency: Strengthening compliance, transparency, and customer confidence through responsible AI


 

•    Beyond Pilots: From experiments to enterprise-wide adoption.
•    Scaling with Trust: Governance, compliance & risk built in.
•    Measurable Impact: Revenue, efficiency & customer value.
•    Future-Proofing Finance: AI ready for shifting regs & markets

•    Building a financial sector-specific AI strategy
•    Integrating AI with existing business processes and systems
•    Establishing key performance indicators (KPIs) for AI adoption

Eighty-five percent of leaders say adapting their workforce to AI is critical. Seven percent are doing it. David Katz led the deployment of frontier AI across BlackRock's active investment business and beyond - starting not with the easiest users, but the most demanding. He shares a counterintuitive measurement approach that revealed where real value lives, the adoption engine that scaled it from hundreds to thousands, and why he thinks that firms that invest in their people will outperform those that replace them